The present invention relates to differential methods for carrying out ion mobility spectrometry analyses, and in particular to the simultaneous analysis of more impurities contained in a very pure gas, such as those employed in the microelectronic industry.
Ion mobility spectrometry is generally known in the field of chemical analyses with the abbreviation IMS, which also indicates the instrument for carrying out the analytic technique (in this case abbreviating the terms Ion Mobility Spectrometer).
The interest for the IMS technique comes from its extremely high sensitivity, associated with the reduced size and cost of the instrument. By operating in suitable conditions it is possible to detect species in gas or vapor phase in a gas mixture in picograms quantities (pg), i.e., 10−12 grams), or in part per trillion concentrations (ppt), i.e., equal to a molecule of analyzed substance per 1012 molecules of sample gas. The gas forming most of the gas mixture will be referred to hereinafter as “carrier gas,” while the mixture itself will be referred to hereinafter as “sample gas.”
There are many application fields of this technique, both in civilian sectors (in particular, for the detection in the industry of inorganic or organic contaminants in clean rooms or of noxious species in the industrial exhausts) and in military sectors (in particular, for the detection of the presence of explosive or toxic substances, such as nerve gases). IMS analysis methods and instruments are disclosed, for example, in U.S. Pat. Nos. 5,457,316, (Cohen et al.); 5,955,886 (Cohen et al.); and 6,229,143 (Werlund).
The general geometrical structure of an IMS instrument is shown in the only attached drawing. The instrument is essentially made up of a cylindrical chamber C having at one end the inlet IC for the sample gas, around which an inlet electrode ER and a ionizing member IM are arranged, and at the opposite end a detector D for charged particles. Detector D is normally kept at the ground potential, while the inlet electrode ER is kept at a potential higher than that of the detector (instrument working in the positive mode) or lower (negative mode) than said detector potential. For simplicity and clarity, in the remainder of the description, reference will always be made to the use of an IMS instrument in the positive mode (the ions corresponding to the analyzed species will be indicated as positive ions), which corresponds to the most common condition of use, but all these considerations are also valid for the use in the negative mode. An electrifiable grid G divides chamber C into two zones, referred to in the field as the “reaction zone” (i.e., on the inlet side of the instrument) indicated in the following with RZ and the “drift zone” (i.e. on the detector side of the instrument) indicated in the flowing with DZ. Grid G generally comprises two alternated series of linear and parallel conductive members, commonly metal wires, which are arranged perpendicularly to the longitudinal axis x of chamber C. A series of electrodes Ei(i=1, . . . , n), generally annular, each set at a suitable voltage for creating between the inlet electrode ER and detector D an electric field suitable for transporting the ions toward said detector, is arranged along the walls of the two zones RZ and DZ. A screen grid SG is arranged between the last electrode En and detector D to avoid the creation on the latter of an “image charge,” (i.e., the build up of an electrostatically induced charge on detector D. The wires of grid G are inserted into the central cavity of an electrode Eg, referred to hereinafter as grid electrode, from which are however electrically insulated. During their motion, the ions are slowed down by a gas usually counter-flowing with respect to the ion motion direction. This gas is introduced from a duct DC at the end of chamber C where detector D is located and is expelled by an outlet OC at the opposite end. The counter-flowing gas, defined as “drift gas” in the field, is an ultra-pure gas which can be the same as the carrier gas or different.
The ionizing member IM is commonly a beta radiation source comprising 63Ni. According to the working mode, the ions with a charge opposite to the charge of the inlet electrode ER are neutralized on or near the inlet electrode ER, while the ions with a charge of the same sign as this electrode undergo a repulsion and are accelerated in the reaction zone. The first ionization produces essentially exclusively ions of the carrier gas, due to its concentration being higher by several orders of magnitude than that of the other species, generally present as traces. The primary ions corresponding to the carrier gas are called “reactant ions” in the field. In the reaction zone, the charge of the reactant ions is distributed among the present species according to their electron or proton affinities, to their ionization potentials or to their electronegativity, according to reactions of the kind:R++Si→R+Si+  (I)wherein R+ represents a reactant ion, R a neutral molecule deriving from the neutralization of the reactant ion (that is, a carrier gas molecule), Si(i=1, . . . , n) a molecule of the i-th species to be analyzed and Si+ the ion corresponding to Si. The ions Si+ often give rise to complex species due to the association with one or more neutral molecules, but for ease of notation and without losing generality, reference will always be made to simple ions hereinafter.
All these ions are transported by the electric field toward the electrifiable grid G. The grid members are grouped in two mutually alternated series, so that each member of one series has two members of the other series as the closest members. The two series of grid members are normally biased with potential values higher and lower, respectively, than the potential of the grid electrode Eg. A transversal electric field generally stronger than the one along the longitudinal axis of the chamber at that point is thus created on the grid plane, so that the ions present in the reaction zone are accelerated toward the members of one of the two series constituting the grid and neutralized. In these conditions, the grid is “closed” and prevents the ions from passing toward the drift zone. When the analysis is to be carried out, the two series of grid members are brought at the same potential of the grid electrode, thereby canceling the transversal field. In these conditions, the grid is “open” so that the ions can advance into the drift zone. The grid opening lasts generally from some tens to some hundreds of microseconds (ms) and during this time a portion of the ions present in the reaction zone is transferred to the drift zone. In particular, the grid is crossed by the ions contained in a cylindrical volume in the reaction zone adjacent to the grid, the height of which is determined by the relation:la=vi×ta  (II)where la is the height of the cylindrical volume, vi is the motion speed of the ion Si+ and ta is the opening time of the grid. The initial or central instant of the grid opening time slot is commonly assumed as the “time zero,” i.e., the analysis start.
In the drift zone DZ, the ions are transported toward detector D with a motion speed which is the resultant of the acceleration due to the presence of the axial electric field and the deceleration due to the collisions with the drift gas. In particular, the motion speed of the i-th ion depends linearly on the electric field and is directly proportional to the temperature T and inversely proportional to the pressure P, according to the effect that the latter two terms have on the viscosity of the drift gas. While the acceleration due to the electric field acts to the same extent on the ions having the same charge (but in the IMS all the ions generally have a unitary charge), the deceleration acts in a different way on the ions according to the different size, shape and mass of the same, so that each ion has a characteristic motion speed and therefore a crossing time of the drift zone (defined as “drift time” in the field) generally different from that of the other ions. By recording the charges collected on detector D, a spectrum is obtained comprising ion current peaks as a function of the time elapsed from the test start. The intensity of each peak in the spectrum is proportional to the amount of charge CSi+ transported by the ion Si+ which caused the peak.
Through calibrating tests, in which sample gases containing a single species Si are analyzed, it is possible, in principle, to obtain data such as speed and drift time of the ionic species Si+ in a given gas and at given temperature and pressure conditions, as well as the efficiency of the reaction I for that species. In ideal conditions and operating in the same conditions of the calibrating tests, the data could be employed in an IMS analysis for determining the presence of a species Si in the gas under exam according to the position of the peaks in the spectrum and its concentration according to the relative size of the different peaks.
However, in the real analyses, the situation is much more complex, due to many phenomena which affect the above theoretical conditions.
A first phenomenon includes the possible presence of unexpected and unknown species Ui(i=1, . . . , m), for which calibration data are not available and that may interfere with the analysis by subtracting charge from the ions Si+ or from the ions R+ according to reactions of the kind:Si++Ui→Si+Ui+  (III)
The result is a spectrum in which the peaks relating to the ions Si+ and the peak relating to the ion R+ (defined in the field “reactant ion peak”, or with its abbreviation RIP, which will be used hereiniafter) have an intensity that is lower than in theoretical conditions or may even disappear, while there are peaks which cannot have an attribution.
Furthermore, the species formed in the reaction zone may react with other neutral species, already in the reaction zone or in the drift zone, with reactions of the kind:Si++Sj→Si+Sj+  (IV)orSi++Si→(Si)2+  (V)Each reaction (IV), (V) proceeds to different degrees and at different rates according to the different kinetics and equilibrium constants for each reaction. These reactions cause the modification of the concentrations of the ions reaching the detector of the IMS instrument with respect to the concentrations initially formed by the direct reaction with the R+ ions, so that the ions corresponding to a species could completely disappear and the latter cannot be detected anymore in the analysis. Reference can be made to the book “Ion Mobility Spectrometry”, edited by G. A. Eiceman and Z. Karpas, published in 1994 by CRC Press, for a presentation of the complex charge transfer principles involved in these reactions.
Furthermore, gases like O2, CO, H2, H2O, etc., coming for instance from previous analyses, can be present on the inner surfaces of the instrument (such as the inner walls of the chamber, electrodes, etc.), either chemisorbed or physically adsorbed. Alternatively, these gases can be dissolved in the materials constituting the instrument. For example, steel, which the chamber is generally made of, normally contains hydrogen. These gases are released both in the reaction zone and in the drift zone. During the analysis, they form additional species Bi(i=1, . . . , r) which come into reactions of type I, III or IV. When these reactions occur in the reaction zone, a charge is removed from the species initially present in the sample gas and spurious peaks appear in the final spectrum. The same reactions, when they occur in the drift zone, may instead lead to spectrum distortions. Unlike the ions of the species Si, which come into the drift zone all at the same time and start from the same position (the grid), the ions corresponding to these species are formed at different points of the drift zone, and therefore, reach the detector at different times according to the formation point: the consequence is that between the spectrum peaks the baseline is not zero as it should be in theory, but there is instead a non-null spectrum “background” which complicates the determination of the area of the peaks or may make it practically impossible in the case of peaks with a lower intensity. The presence of the species Bi also involves other drawbacks. First, these may react with species Si+ through reactions of type III or IV in the drift zone, thus causing an undesired attenuation of the charge quantity transferred by the species Si+ to the detector and a consequent reduction of the instrument sensitivity. Second, the interaction of species Bi with the species R+ in the drift zone may cause an undesired attenuation of the RIP, with a consequent reduction of the upper reading threshold of the instrument.